Typically, the execution of a critical enterprise application has to meet a set of service level agreements between a software company that provides the application and the enterprise that uses the application. For example, the service level agreement for a healthcare application can require that processing an application operation to schedule a doctor's office visit should not take more than an average of 3.0 seconds, and require that processing an application operation to order medication should not take more than an average of 10.0 seconds. In another example, the service level agreement for a banking application can require that processing an application operation to approve a loan should not take more than an average of 5.0 seconds, and require that processing an application operation to produce a quarterly profits report should not take more than an average of 10.0 minutes. If the execution of an enterprise application's operations does not meet its service level agreements, the enterprise may not be efficient enough, and may lose millions of dollars by not being able to process as many application operations as expected, such as serving mobile phone calls or providing loans. A software company that provides the application may be required to pay fees to the enterprise when the processing of an application's operations does not meet the service level agreements. Preventing the violation of a service level agreement or resolving such a violation quickly requires the identification of the root cause of an application underachieving a service level agreement, but such identification is not an easy task, particularly in a virtually provisioned environment. A virtually provisioned environment may use thousands of servers in a data center to process application operations, and some application operations may be processed by dozens of servers. Identifying the root cause of application underachievement in a virtually provisioned environment may be quite challenging due to the vast number of server resources involved in processing application operations. Even after the root cause is identified for application underachievement in a virtually provisioned environment, efficient adjustments to address the root cause may be challenging for the combination of thousands of servers and hundreds of service level agreement requirements.